Aspen HYSYS Simulation of CO2 Removal by Amine Absorption from a Gas Based Power Plant

SIMS2007 Conference, Gøteborg, October 30-31st 2007

Lars Erik Øi

Telemark University College, Norway lars.oi@hit.no

Abstract A simplified combined cycle gas power plant and a MEA (monoethanol amine) based CO2 removal process have been simulated with the process simu-lation tool Aspen HYSYS. The thermodynamic prop-erties are calculated with the Peng Robinson and Amines Property Package models which are avail-able in Aspen HYSYS. The adiabatic efficiencies in compressors, gas turbines and steam turbines have been fitted to achieve a total thermal efficiency of 58 % in the natural gas based power plant without CO2 removal. The efficiency is reduced to about 50 % with CO2 removal. The CO2 removal in % and the energy consumption in the CO2 removal plant are calculated as a function of amine circulation rate, absorption column height, absorption temperature and steam temperature. With CO2 removal of 85 %, heat consumption is calculated to 3.7 MJ/kg CO2 removed, close to a literature value of 4.0 MJ/kg CO2.

Keywords: Aspen HYSYS, CO2 removal, mono-ethanol amine, absorption.

1. Introduction

Figure 1. Principle for CO2 removal process based on absorption in amine solution.

The possibility of removing CO2 from fossil fuel based power plants has got increased interest due to environmental reasons. The most actual method for CO2 removal is by absorption in an amine based sol-

vent followed by desorption. The principle of the removal process is shown in figure 1. The simplest and most used amine for CO2 removal is MEA (monoethanol amine). This removal process has a high consumption of thermal energy, and one of the main aims of improvement is to minimize this en-ergy consumption. In the case of CO2 removal from a combined cycle gas power plant, it is natural to cover the heat requirement for CO2 stripping with steam from the power plant. A possible steam outlet from a typical combined cycle process is shown in figure 2.

Figure 2. Principle of combicycle power plant

Because testing at large scale is so expensive, it is natural to use process simulation to evaluate such processes. There are however few literature refer-ences on process simulation of CO2 removal from exhaust gases at atmospheric pressure. There are some journal articles [1-4], and 3 of them have used the process simulation program Aspen Plus. As-penTech bought the program HYSYS from Hypro-Tech in 2002, and in 2006 the program name was changed to Aspen HYSYS. The last reference [4] uses a Fortran code to simulate the process. An im-

73

portant advantage with using a process simulation program for such calculations, is that the available models for thermodynamic properties can be used. Aspen Plus has an MEA property insert model [5] which was used in all the references [1-3]. Aspen HYSYS has an Amines Property Package [6]. Within the Amines Property Package, one of the two models, Kent Eisenberg or Li-Mather, can be se-lected. Even though Aspen HYSYS is probably the most used process simulation program in the world, there has not been found any journal references on CO2 removal from atmospheric exhaust gas using this program. There is however much literature on CO2 removal from natural gas at high pressure. Other process simulation programs containing amine packages, are ProVision and ProMax.

At Telemark University College, Hysys has been used since 2000 to simulate CO2 removal from gas based power plants. Most of this work has been Bachelor and Master student projects with the author of this paper as supervisor. Major challenges in the simulation of CO2 absorption and desorption proc-esses, are the description of thermodynamics and ab-sorption efficiency, convergence and total energy or cost optimization.

The purpose of this paper is to present and discuss the use of Aspen HYSYS for simulating CO2 re-moval from atmospheric exhaust by amine absorp-tion, and to simulate the total effect of efficiency re-duction in a combined cycle gas power plant.

2. Available models

2.1 Absorption and reaction mechanisms

The details of the mechanisms of CO2 absorption into an amine solution in an absorption column are quite complex. There are many references about the chemistry involved in the process, and many refer-ences and models comprising mass transfer mecha-nisms and chemical reaction kinetics. Reviews are written by Danckwerts and Sharma [7] and Versteeg [8].

First, CO2 has to be transported from the gas to the liquid surface, and then it is absorbed in the liquid solution. The gas liquid interface area a (in m2/m3) and liquid holdup h (in m3/m3) are main parameters in describing such mechanisms.

In the liquid, the CO2 may react chemically with other components. The following reactions are nor-mally assumed to take place when CO2 reacts in a primary amine like MEA in an aqueous solution. In the case of MEA (NH2C2H2OH), R is C2H2OH.

CO2 + NRH2 → RH2+NCOO- (1)

RH2+NCOO- + NRH2 → RH2NCOO-NRH2

+ (2)

According to equation (1) and (2), two moles of MEA are necessary to absorb one mole of CO2.

A simple overall description of the combined ab-sorption and reaction process is simply

CO2 (gas) → CO2 (absorbed) (3)

The removal of CO2 is not 100 %. The % CO2 re-moval is limited both by low absorption and reaction rates and by the equilibrium conditions.

If the kinetics in the reactions should be calculated, more details about the intermediate reactions in equation (1) and (2) should be included. This is done in the MEA property insert model in Aspen Plus.

The simulation program Aspen HYSYS is mainly based on equilibrium calculations. In that case, equation (3) is sufficient to calculate the absorption process.

2.2 Equilibrium models

The concentration of CO2 in the gas may be ex-pressed by the partial pressure pCO2 (eg. in bar) and the concentration in the liquid may be expressed by CCO2 (eg. in mole/m3).

The equilibrium between the concentrations of CO2 in a gas and a liquid may be represented as a func-tion

pCO2 = f(CCO2) (4)

This expression might be a function of temperature, pressure and concentrations of the components in the solution. There are many models available to de-scribe this function.

In Aspen HYSYS, gas/liquid equilibrium for a com-ponent (i) is normally calculated using k-values de-fined by the equation

Ki = yi/xi (5)

where yi and xi are the mole fractions of (i) in the gas and liquid phase. For general purpose use, equa-tion of state models like SRK (Soave Redlich Kwong) and PR (Peng Robinson) are often used. Aspen HYSYS recommends Peng Robinson [9]. Peng Robinson is regarded to be suitable to handle systems containing hydrocarbons, water, air and combustion gases, the typical components in a natu-ral gas based power plant.

Traditional equation of state models are not regarded to be suitable for non-ideal liquid systems. An amine solution is an electrolytic system and also

74

comprises chemical reactions. This is not expected to be well described with traditional equations of state.

Within the Amines Property Package in Aspen HY-SYS, one of the two models, Kent Eisenberg [10] or Li-Mather [11], can be selected. The models are quite complex, but in principle they are models to describe the equilibrium of the CO2 concentrations in the gas and the liquid (equation 4).

Aspen Plus has an electrolytic package to calculate liquid systems containing ions. Different electrolyte equilibrium models can be used. Using an MEA property insert model, equilibrium models can be combined with reaction kinetic models, including rate expressions of chemical reactions like equation (1) and (2).

2.3 Column models in simulation programs

A CO2 absorption column is a unit where gas flows up and liquid (eg. an amine solution) flows down. CO2 is transfered from the gas phase to the liquid phase where it reacts with the amine solution. The gas and liquid phases are made to get in contact by the help of column plates or random or structured packing.

The CO2 stripping column also has plates or pack-ing, and this column also has a reboiler at the bottom to provide heating, and a condenser at the top to pro-vide cooling.

The traditional way to model such columns is by us-ing equilibrium stages. One plate can be calculated assuming equilibrium between the CO2 concentra-tion in the gas and liquid leaving the plate. In a packed column, a certain height of packing can be modeled as one equilibrium stage.

The equilibrium stage model can be refined by in-troducing a stage efficiency. Murphree efficiency for stage number n is defined by

EM = (y-yn-1) / (y*-yn-1) (6)

where y is the mole fraction of CO2 in the gas leav-ing the stage, yn-1 is the mole fraction leaving the stage below, and y* is the mole fraction CO2 in equi-librium with the liquid leaving the stage. This is il-lustrated in figure 3.

Figure 3: Definition of Murphree efficiency

Most process simulation programs have models for implementing Murphree efficiency in a column model. The Amines property package in Aspen HYSYS has a special estimation method for predict-ing this Murphree efficiency. This is based on the work of Tomkej [12]. This model is based on ex-perience with CO2 removal from high pressure natu-ral gas. In Aspen Plus it is also possible to specify these efficiencies explicitly in an absorption or dis-tillation column model.

Aspen Plus has a column model (RateFrac) which can include kinetic rate expressions in chemical re-actions at each stage. In the references [1-3] the MEA property insert model in Aspen Plus is used in a column model.

2.4 Column convergence

To converge a column model in a process simulation program, all the equations describing the equilibrium and gas and liquid flows must be solved for each calculation stage. Including rate expressions for chemical reactions complicates the calculations fur-ther. This often leads to convergence problems.

The column model in Aspen HYSYS has a default set of convergence criterias, and a default set of cal-culation parameters. Different calculation models are also available. The Inside-Out algorithm is de-fault, and a Modified Hysim Inside-Out algorithm is also available. A damping parameter for the column iteration is adjustable, and the damping can be speci-fied to be adaptive.

75

2.5 Flowsheet convergence

All process simulation programs are based on mod-ules for calculating different unit operations like heat exchangers, pumps, distillation columns etc. Process simulation programs are traditionally divided into ei-ther sequential modular or equation based programs. In a sequential modular program, the in-streams of each calculation module must be known prior to the calculation, and the out-streams are the result of the module calculation. The programs Aspen Plus and ProVision are sequential modular. Equation based simulation programs can be able to calculate in-streams based on out-streams. Aspen HYSYS is an equation based simulation program.

However, also in Aspen HYSYS, the column models are based on specified in-streams. Because of this, flowsheets with columns in practice have to be cal-culated in a modular sequential manner.

In many cases, it is of interest to calculate in-streams also to columns. This can be done by iteration methods. In the case of recycle streams, the flow-sheet can be solved including recycle blocks. A re-cycle block compares the in-stream to the block with the out-stream from the block in the former iteration.

In the case of convergence problems in a column model, recycle iterations complicate the calculations further. In some cases, a recycle block will not con-verge due to parameters of minor interest. An ex-ample of such a parameter is the concentration of a trace component. In such cases, a possibility is to it-erate manually on the main parameter (eg. the CO2 concentration) by replacement, and accept the errors in the parameters of minor importance.

3. Calculations

3.1 Power plant simulation

Figure 2 shows the principle of a combined cycle power plant based on combustion of natural gas. A real plant will be much more complex with many heat exchangers, recycle pipes and tanks to optimize the overall efficiency of the power plant. However, a process as in figure 2 is practically realistic, but it will not have an optimum overall efficiency.

The main purpose of the power plant simulation is to make a simple but realistic model to evaluate the influence of heat for CO2 removal on the overall power plant efficiency.

A simplified combined cycle 400 MW gas based power plant has been simulated with Aspen HYSYS. Pure methane is used as natural gas, air is 79 % ni-trogen and 21 % oksygen, 100 % combustion is as-sumed, and traditional temperatures and pressures are used in the process. The Peng Robinson model has been used for the thermodynamic properties in the power plant. Specifications for the calculation are listed in table 1. A flowsheet of the Aspen HY-SYS model is presented in figure 4. This model has been developed during the Master Thesis work of Kristin Vamraak [13] and Bjørn Moholt [14]. Inlet air/gas temperatures 25 ºC Inlet natural gas pressure 30 bar(a) Combustion temperature 1500 ºC Steam high pressure 120 bar(a) Steam medium pressure 3.5 bar(a) Steam low pressure 0.07 bar(a) Pressure to stack 1.01 bar(a)

Stack temperature 100 ºC Table 1: Specifications for Power Plant Simulation

Figure 4. Aspen HYSYS model of simplified gas power plant.

76

To have a physically possible prosess, the flue gas temperature (680 ºC from gas turbine to exhaust at 100 ºC) has to be higher than the steam temperature through all the steam heat exchanger. This results in a maximum superheated steam temperature of 540 ºC (at 120 bar). The chosen pressure is not opti-mized.

The efficiencies in compressors, gas turbines and steam turbines have been fitted to achieve a total thermal efficiency of 58 % (which is traditional) in the gas based power plant. The total efficiency is calculated as the turbine effects (minus compressor and pump) divided by the lower heating value of natural gas. The compressor efficiency (adiabatic) was adjusted to 90 %, the expander part of the gas turbine (expander) and the steam turbines were ad-justed to 85 % (also adiabatically). These values are high compared to actual efficiencies for such equip-ment. A real power plant would be more energy op-timized, and have lower equipment efficiencies.

In this part of the power plant simulation, the steam delivery (for CO2 removal) is set to zero.

3.2 Simulation of CO2 removal base case

An absorption and desorption process for CO2 re-moval with an aqueous MEA solution has been simulated. The exhaust gas from the power plant model is used as the feed to this model. The absorp-tion column is specified with 10 stages each with a Murphree efficiency of 0.25. (An estimated HETP (Height Equivalent to a Theoretical plate) of 4 me-ter, is about equivalent to 0.25 efficiency for each meter of packing.) Traditional concentrations, tem-peratures and pressures are used in the base case simulation. The thermodynamics for this mixture is

described by an Amines Property Package available in Aspen HYSYS. The Kent Eisenberg [10] model is selected in the Amines Property Package. Specifications for the calculation are listed in table 2. The Aspen HYSYS CO2 removal model is pre-sented in figure 5. Different versions of this model have been developed in several student projects. The version in figure 5 is based on a Master Project work by Trine Amundsen [15]. Inlet gas temperature 40 ºC Inlet gas pressure 1.1 bar(a) Inlet gas flow 85000 kmole/h CO2 in inlet gas 3.73 mole-% Water in inlet gas 6.71 mole-% Lean amine temperature 40 ºC Lean amine pressure 1.1 bar(a) Lean amine rate 120 000 kmole/h*) MEA content in lean amine 29 mass-% *) CO2 in lean amine 5.5 mass-% *) Number of stages in absorber 10 Murphree efficiency in absorber 0.25 Rich amine pump pressure 2 bar Heated rich amine temperature 104.5 ºC *) Number of stages in stripper 6 (3 + 3) Murphree efficiency in stripper 1.0 Reflux ratio in stripper 0.3 Reboiler temperature 120 ºC Lean amine pump pressure 2 bar Minimum deltaT in heat exch. 10 ºC *) In first iteration Table 2: Specifications for Base Case CO2 removal

Figure 5. Aspen HYSYS model of CO2 removal.

77

85 % CO2 removal can be specified in the process. The Kent Eisenberg equilibrium model has been compared with the Li-Mather equilibrium model [11]. The CO2 removal calculated by Aspen HYSYS was reduced from 85 to 82 %, and the heat con-sumption was reduced from 3.65 to 3.4 MJ/kg CO2.

3.3 CO2 removal sensitivity calculations

3.3.1 Variables held constant

The model has been used to evaluate the effects of changing the most important parameters.

In most of the calculations, the CO2 removal and the stripping heat consumption were calculated, while keeping all the other parameters in table 2 constant. From a calculation viewpoint, this is probably the simplest.

Another possibility had been to keep the % CO2 re-moval constant, and calculate the heat duty and the necessary column height. This would give the pos-sibility to optimize the trade-off between operation cost (due to heat consumption) and capital cost (due to column height).

In the cases where the default Inside-Out algorithm did not converge, the Modified Hysim Inside-Out algorithm with adaptive damping was tried to obtain convergence.

3.3.2 Circulation rate

The effect of increased circulation rate, is that the removal grade increases. The results of the simula-tions are shown in figure 6. A minimum calculated steam consumption is calculated to 3.62 MJ/kg CO2 removed.

Figure 6: Circulation rate dependence

3.3.3 Number of absorption stages

The height of the absorption column is varied by changing the number of stages. The Murphree stage

efficiency for CO2 is kept constant at 0.25. The height can also be changed by varying the stage effi-ciency. As expected, removal grade increases and heat requirement decreases with increased column height. The result is shown in figure 7. The calcula-tion diverges using more than 12 stages in the col-umn.

Hysys has calculated estimated EM (equation 6) for a plate, and the resulting EM for a plate varied between 0.08 and 0.13.

Figure 7: Number of stages dependence

3.3.4 Absorption temperature

An increase in gas and liquid inlet temperature leads to reduced absorption at equilibrium. Simulation re-sults based on a constant stage efficiency are shown in figure 8. In practice, a higher temperature will give a higher absorption and reaction rate, but the equilibrium results will not show this effect.

Figure 8: Absorber temperature dependence

78

3.3.5 Absorption pressure

The absorption pressure is set to atmospheric pres-sure at the outlet and atmospheric pressure plus pres-sure drop at the inlet. The pressure inlet in the base case was 1.1 bar. In the case of a pressure drop of 0.5 bar, the % CO2 removal increased to 87.1 % and the energy consumption was reduced to 3.59 MJ/kg CO2 removed.

3.3.6 Reboiler temperature

Increased reboiler temperature gives purer amine so-lution and better CO2 removal efficiency. However, amine degradation problems arise above 120 ºC. The temperature was varied between 118 and 121 ºC. It was difficult to get converged solutions out-side this range. At 121 ºC, an outside range warning was given. The results up to 120 ºC are shown in figure 9.

Figure 9: Reboiler temperature dependence

3.3.7 Stripper pressure

The stripper pressure was specified to 2 bar(a) in the calculation. It was very difficult to get a con-verged solution at other pressures. A solution with a warning (outside range) was achieved with a pressure at 1.9 bar(a).

3.3.8 Minimization of heat consumption

There have been performed many Aspen HYSYS calculations at different conditions. One aim is to reduce the heat duty as much as possible. The low-est heat consumption calculated was 3.39 MW/kg CO2 removed. The CO2 removal efficiency was then 93.8 %. This was obtained with gas tempera-ture at 30 ºC, 16 absorption stages, stripper pressure 2.0 bar and reboiler temperature at 200 ºC. At lower temperatures or more absorption stages, the calcula-tion did not converge.

3.4 Power plant efficiency reduction due to CO2 removal

The steam consumption in the CO2 removal process is delivered from the gas power plant as shown in figure 2, and the total thermal efficiency is reduced. Steam at 3.5 bar is delivered at 139 ºC. When the base case heat consumption was used, the total effi-ciency was reduced from 58 % to 53 %. If the effi-ciency loss due to exhaust gas fans and circulation pumps was included, the resulting total efficiency would be approximately 50 %. If energy for CO2 compression and condensing was to be delivered from the power plant, the total efficiency would be below 50 %.

The calculated alternative with the lowest duty was with 3.39 MJ/kg CO2 removed. This gives a re-duced reduction in total efficiency of 0.4 % (%-points).

The effect of a possible lower temperature for heat stripping can be calculated. The medium steam pressure can be reduced slightly. The same duty is used, and the total effect should give a slight in-crease. This effect is so marginal, that it is within the uncertainty of the calculations.

4. Discussion and Conclusion

4.1 Accuracy

With the same specifications, the calculated results vary slightly dependent on initial values. The accu-racy in the calculations is normally within 1 % (ab-solute) in the CO2 removal and a few percent (abso-lute) in heat consumption. This accuracy can proba-bly be improved with tighter convergence limits.

The uncertainty due to equilibrium is probably higher. This is indicated by the CO2 removal chang-ing with 3 % (%-points) when changing equilibrium model from Kent Eisenberg to Li-Mather. The un-certainty in the power plant calculation due to the Peng Robinson equation of state has not been checked, but is expected to be less than the uncer-tainty in the amine models.

4.2 Evaluation of the simplified process

The calculated CO2 removal process is a simplified process. Heat losses and some pressure losses are neglected. A real process contains more equipment, pipes and valves, and all this equipment also have heat losses and pressure losses.

A real process with MEA will probably include a water wash section to reduce MEA emissions and a reclaimer unit to recover MEA from thermally de-

79

generated MEA. These additional units will in-crease the energy consumption.

There are possibilities to reduce the energy con-sumption in the process by different stream configu-rations (eg. split stream). These possibilities will however normally increase the investment.

The calculated heat consumption of 3.65 MJ/kg CO2 is regarded as realistic. This is slightly lower than normally found in literature, eg. Desideri [1] which has a list of references with values mostly above 4.0 MJ/kg CO2. If water wash is included in the highest value, this can explain the difference. The lowest calculated value of 3.39 MJ/kg CO2 is regarded as an optimistic value. It might be regarded as a poten-tial value for improvements. The cost optimum heat consumption will probably be a trade-off between investment and operational cost.

4.3 Convergence problems

There are many problems with convergence in these calculations. The problems normally occur in the absorption or stripping column. One problem is numerical. It is found that the Modified Hysim In-side-out algorithm with adaptive damping gives the best convergence.

If there are too many stages specified in the col-umns, they tend to diverge. That is traditional for column stage calculations in typical process simula-tion tools.

There is also a problem with the range for the model. The Kent Eisenberg model is limited to below 30 weight-% MEA and below 120 ºC. The simulation is calculated, but the program gives a warning.

4.4 Parameter variations

Different parameters have been varied to calculate the effects and to make a tool for optimizing the process.

The only parameter that is varied so that an optimum is found, is the recirculation rate. In figure 6 it is shown that a minimum heat consumption (of 3.62 MJ/CO2) is achieved at a lean amine circulation rate of 2550 ton/h.

When the number of stages is increased, the per-formance of the process increases, but the calcula-tion tends to diverge. However, when the column diverges due to too many stages, the % CO2 removal and heat consumption is probably close to a maxi-mum.

A reduction in absorption temperature leads to im-proved performance according to figure 8. Accord-ing to literature, a temperature of about 40 ºC is rec-ommended. At lower temperature, the reaction rates

will be too slow. The calculations have been per-formed with constant column stage efficiencies of 0.25. The original Aspen HYSYS calculates effi-ciencies are temperature dependent. A temperature dependent stage efficiency is probably necessary to find an optimum absorption temperature.

It is not difficult to calculate the effects of absorp-tion pressure drop. There is a trade-off between CO2 removal efficiency and cost of column height and pressure drop.

Increased reboiler temperature gives purer amine so-lution and better performance. The optimum tem-perature is a trade-off between improved CO2 re-moval and degradation of the amine solution. Aspen HYSYS limits the calculation to 120 ºC as in the base case. The resulting lean loading (mole CO2/mole amine) in the liquid to the absorption col-umn is 0.26. This is close to 0.25 as reported as an optimum by Alie [3].

The stripper pressure is an adjustable parameter. It is difficult to get a converged calculation with a stripper pressure far away from 2 bar with 120 ºC. According to Freguia [2], the pressure should be be-tween 1.5 and 2 bar. It is not obvious whether the convergence difficulties are due to physical limita-tions, or if it is a numerical problem. The pressure must be consistent with the pressure of the amine so-lution leaving the reboiler. The problems of limited range of reboiler temperatures and stripper pressures are probably related.

4.5 Aspen HYSYS compared to other tools

There are other tools to simulate such processes. Other commercial process simulation tools like ProVision, ProMax or Aspen Plus can be used to simulate the process in a similar way. Aspen Plus also has the possibility to calculate rate expressions on an ideal mixing stage (simulating a column plate). This has the advantage of taking into account the reaction rates for different reactions simultane-ously. It is possible (but difficult) to include rate expressions of absorption (transport of CO2 from the gas to the liquid phase) in such a model. It is also a question whether this kind of a mixing stage model is a good model for continuous countercurrent op-eration as in structured packing.

The presented Aspen HYSYS model is based on a specified Murphree efficiency for each stage (or height of packing). It is possible to make this effi-ciency a function of rate expressions for the absorp-tion rate and the reaction rates.

It is of course possible to simulate CO2 removal processes without using commercial process simula-tion programs. It is however necessary to include at

80

least one reliable and robust equilibrium calculation model and one robust column model. It is difficult to compete with the commercial process simulation programs in these two matters. The commercial programs also normally have very good input and output facilities.

A problem with the commercial process simulation programs, at least from an academic point of view, is that some of the models of interest are not docu-mented accurately.

4.6 Further development

There are available calculation models for the design of CO2 removal plants based on MEA. The largest uncertainty is probably connected to the absorber stage efficiency. In the case of a structured packing column, this gives a large uncertainty in the neces-sary packing height. An improvement here will be of great value.

In the case of other amines than MEA, the limitation is in the equilibrium model or in the uncertainty in the equilibrium model parameters.

In the case of mixed amines, the importance of the reaction rates increases. Such models are probably best treated in a rate based model like RateFrac in Aspen Plus. It is however possible that a practical approach when using Aspen HYSYS, is to perform estimation of Murphree efficiency outside the proc-ess simulation tool.

4.7 Conclusion

The CO2 removal model developed in Aspen HY-SYS is useful for evaluating the effects of changing amine circulation rate, absorption column height, absorption temperature and reboiler temperature. The reduced total efficiency in a gas based power plant can be calculated.

The developed models can be developed further to improve amine absorption processes for CO2 re-moval. The combination of one model for a power plant and another for CO2 removal can be useful for total energy optimization and total cost optimization.

5. Acknowledgements These models are the result of much trial and error work done by my students during many years. Thanks to all of you!

References [1] Desideri, U., Alberto, P. (1999). Performance mod-elling of a carbon dioxide removal system for power plants, Energy Conversion & Management 40, 1899-1915. [2] Freguia, S., Rochelle, G.T. (2003). Modelling of CO2 Capture by Aqueous Monoethanolamine, AIChE Journal, 49, 1676-1686. [3] Alie, C., Backham, P., Croiset, E., Douglas, P.L. (2005). Simulation of CO2 capture using MEA scrub-bing: a flowsheet decomposition method, Energy Con-version & Management 46, 475-487. [4] Tobiesen, F.A., Svendsen, H.F., Hoff, K.A. (2005). Desorber energy consumption amine based absorption plants. International Journal of Green Energy, 2, 201-215. [5] Aspen Plus, AspenTech, Cambridge Mass., 2006. [6] Aspen HYSYS 2004.2, AspenTech, Calgary 2007. [7] Danckwerts, P.V., Sharma, M.M. (1966). The ab-sorption of Carbon Dioxide into solutions of alkalis and amines, The Chemical Engineer, oct. 244-280. [8] Versteeg, G.F., Van Dijck, L.A.J., Van Swaaij, W.P.M. (1996). On the kinetics between CO2 and al-kanolamines both in aqueous and non-aqueous solu-tions. An overview, Chem. Eng. Comm. 144, 113-158. [9] Peng, D., Robinson, D.B. (1976). A New Two-Constant Equation of State, Ind. Eng. Chem. Fundam., 15 No.1, 59-64. [10] Kent, R.L., Eisenberg, B. (1976). Better data for Amine Treating, Hydrocarbon Processing, 55, No.2, 87-90. [11] Lee, I.J., Otto, F.D., Mather, A.E. (1975). Solubil-ity of Mixtures of Carbon Dioxide and Hydrogen Sul-fide in 5.0 N Monoethanolamine Solution, J. Chem. Eng. Data, 20, 161-163. [12] Tomkej, R.A., Otto, F.D. (1986). Improved De-sign of Amine Treating Units by Simulation using Per-sonal Computers, World Congress III of Chemical En-gineering, Tokyo, Sept. 21-25. [13] Vamraak, K. (2004). Energiforbruk ved CO2-fjerning fra gasskrafverk, Master Thesis, Telemark University College. [14] Moholt, B. (2005). Simulering av CO2-fjerning med aminer, Master Thesis, Telemark University Col-lege. [15] Amundsen, T. (2007). CO2-renseanlegg i Aspen HYSYS, Master Project, Telemark University College.

81